TU Dortmund -- Computer Science VIII

7. Fachgruppentreffen Maschinelles Lernen, 1995 und AAAI Spring Symposium on Information Gathering, 6 - 12 G. Kappel, B. Pröll, W. Retschitzegger, W. Schwinger (2001) Customisation for Ubiquitous Web A [...] objectionable web content by leveraging multiple information sources, SIGKDD explorations, 8, 17-26 Ad 6) M. Charikar, C. Chekuri, T. Feder, R. Motwani (1997) Incremental clustering and dynamic information …

doi:10.1016/j.sysarc.2006.02.005

for Fixed Telecomm Networ Networking Community Swarm Intelligence MP-Scout (6.2) OSPF (6.2) MDVA (6.2) MPDA (6.2) M-Path (6.2) Packet switched Packet switched Circuit switched ABC (3.2) ABC-Smart (3.2) [...] with original ABC. 6 4 5 8 9 3 2 7 1 1 1 2 1 1 1 1 2 1 1 Routing Table at Node 9 4 5 6 8 1 p41 p51 p61 p81 2 p42 p52 p62 p82 3 p43 p53 p63 p83 4 p44 p54 p64 p84 5 p45 p55 p65 p85 6 p46 p56 p66 p86 7 p47 [...] represen- tative nodes 1 and 6 respectively, by following the 4 5 3 2 1 2 1 2 1 Routing Tab 1 p41 2 p42 3 p43 4 p44 5 p55 6 p66 7 p67 8 p88 Rep Node Foraging Region 1 IFZ+IFR 8 7 6 5 4 3 2 1 Fig. 4. Routing …

bled.dvi

and Inductive Logic Programming 2 3 A logical framework 6 3.1 Assumptions in ILP : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : : 6 3.2 A partial Herbrand interpretation : : : : : : : : : [...] is one ground instance and Q is injective over M. 6 k For every clause 0 of (), if 0 k then k 0 Theorem 1 I CWA () j= Proof Assume that I CWA () 6j= P Q and P Q 2 . Without loss of generality assume [...] background theory should not entail the positive examples, B 6 kE + (Prior necessity), 2. and it should be consistent with the examples, B;E ; E + 6 k? (Prior satisability). 3. The background theory and the …

Wissensentdeckung in Datenbanken - Optimierung, Überanpassung

Universität Dortmund Optimierung Overfitting Überanpassung Überanpassung 13 von 15 0 0.2 0.4 0.6 0.8 1 0 0.2 0.4 0.6 0.8 1 Bias Varianz E rr o r Complexity Informatik—Künstliche Intelligenz Computergestützte [...] Stephen Boyd und Lieven Vandenberghe: Convex Optimization) 4 von 15 1.15 1.2 1.25 1.3 1.35 0 0.2 0.4 0.6 0.8 1 f( x ) x log(exp(x)+exp(1-x)) Informatik—Künstliche Intelligenz Computergestützte Statistik Technische [...] bei f ∶ Rn → Rm: Falls K = sup c∈Rn ∥∇f(c)∥2 <∞ , dann ist f Lipschitz stetig mit Konstante K > 0. 6 von 15 Informatik—Künstliche Intelligenz Computergestützte Statistik Technische Universität Dortmund …

Blearn3.dvi

diagonal),s jump(7,6,12,26,diagonal), s jump(7,7,13,26,diagonal) and s jump(8,5,11,26,diagonal), s jump(8,6,12,26,diagonal),s jump(8,7,13,26,diagonal) and s jump(19,5,9,26,parallel),s jump(19,6,10,26,parallel) [...] time interval for which the membership is valid. For trace 7 and sensors 5, 6, and 7 we have sclass(7,5,1,26,right side), sclass(7,6,1,26,right side), and sclass(7,7,1,26,right side). To express the fact that [...] Sensor, Time4, Time5, Grad4) & stable(Trace, Orientation, Sensor, Time5, Time6, Grad5) & ! s jump(Trace, Sensor, Time1, Time6, parallel). The rst learning task should adjust the tolerance function for gradience …

LNAI 5212 - Hierarchical, Parameter-Free Community Discovery

Discovery 183 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 DBLP ! Flat Density % su bg ra ph s 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 CLASSIC ! Flat Density % su bg ra ph s 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 [...] 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 DBLP ! Hierarchical Density % le af su bg ra ph s 0 0.2 0.4 0.6 0.8 1 0 20 40 60 80 100 CLASSIC ! Hierarchical Density % le af su bg ra ph s 0 0.2 0.4 0.6 0.8 1 0 20 [...] Second level Fig. 2. Adjacency matrix view (cf. Figure 1) edge code 1 I1 I2 J2J1 G3 G4 G2 G5 G7 G8 G6 G9 G16 G14 G15 G13 G11 G12 G10 { , }I1 I2 { , }J1 J2 Level 2 G A2,1!( ) G Level 0 Level 1 Fig. 3. Cluster …

3500 4000 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 R un tim e (in s ec on ds ) Frequency threshold (in %) Fig. 2 Runtime experiments carried out on the data set C200T2.5S10 I10K 6.2 Results 6.2.1 Synthetic dataset [...] 1e+06 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 N um be r of fr eq ue nt C S P s Frequency threshold (in %) (a) Number of frequent CSPs 1 10 100 1000 10000 100000 1e+06 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 N um be r of fr [...] Conjunctive sequence patterns 4.1 Support bounding and Möbius inversion 5 Algorithm 6 Experimentations 6.1 Experimental method 6.2 Results 7 Conclusions folder.joboptions << /ASCII85EncodePages false /Allow …

Vorlesung Wissensentdeckung in Datenbanken - SVM -- Textkategorisierung

5 1998 1998 1998 4 16423 6 1998 9999 1998 61 5016 1 1997 1999 1997 33 5016 2 1999 2001 1999 33 5016 3 2001 2001 2001 33 5016 4 2001 2001 2001 33 5016 5 2001 2002 2001 81 5016 6 2002 9999 2001 94 . . . [...] Katharina Morik, Claus Weihs LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund 9.6.2009 1 von 46 LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund Web Mining [...] Verwendung des Modells zur Textklassifikation für zeitgestempelte Daten Beispiel zur Klassifikation 6 von 46 LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund Web Mining Textk …

5DMVsvm3_4p

5 1998 1998 1998 4 16423 6 1998 9999 1998 61 5016 1 1997 1999 1997 33 5016 2 1999 2001 1999 33 5016 3 2001 2001 2001 33 5016 4 2001 2001 2001 33 5016 5 2001 2002 2001 81 5016 6 2002 9999 2001 94 . . . [...] Katharina Morik, Claus Weihs LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund 9.6.2009 1 von 46 LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund Web Mining [...] Verwendung des Modells zur Textklassifikation für zeitgestempelte Daten Beispiel zur Klassifikation 6 von 46 LS 8 Informatik Computergestützte Statistik Technische Universität Dortmund Web Mining Textk …

= ∑ v∈R yv 5 return S, λ 6 Function FairDecomposition(G, L, R) 7 k = 0 8 L′, R′ = L,R 9 while L′ 6= ∅ do 10 X, λ′ = SmallestRatioSet(G, L′, R′) 11 L′ = L′ \X 12 if k = 0 or λ′ 6= λ then /* Create new block [...] b2), (a3, b1)}. The distribution F1 choosing from among M1,M2,M4 and M5 with probability 1 3 , 1 6 , 1 6 , 1 3 , respectively, is maxmin-fair. The satisfaction probabilities of a0, a1, a2 and a3 are then [...] are equivalent: (1) D is maxmin- fair; (2) D is minmax-Pareto; (3) D is Pareto-efficient and D↓ 6= E↓ 6= for all Pareto-efficient distributions E over S. The proof may be found in Appendix A. Note that …